Analyzing proto-s cluster distribution

[venkatachalapathy]

Comparing clusters seem to be equivalent to comparing multinomial distributions. Comparing clusters with unknown number of components seem to involve nonparametric versions of multinomial distribution comparison problem.

[divilian]

(Oh, right, duh! Multinomial distros are of course the right thing as soon as you think about it.)

[venkatachalapathy]

This package uses some fancy metric called Wasserstein metric for comparing two distributions

https://cran.r-project.org/web/packages/twosamples/index.html

This is the first one I could find but there could be others